Keyword Research: A Complete, Repeatable Process

Keyword research is the process of finding what your customers type into Google, sorting those terms by intent and difficulty, and deciding which ones you’ll actually build pages around. Done properly, it takes eight repeatable steps: define the goal, collect seed terms, expand the list with tools, mine competitors, judge intent, check difficulty and volume, cluster keywords into pages, and mine Search Console for quick wins. Most small sites can run this end to end in 15-25 hours the first time, then in 3-5 hours per quarter after that. This guide walks through every step, and links out to a deep-dive post on each one.

What Is Keyword Research, Exactly?

Keyword research is the practice of identifying the exact words and phrases real people search for, then matching those phrases to the intent behind them before you write a single sentence of content. It is not “picking words that sound relevant.” A keyword list without intent labels, difficulty scores, and a mapping to specific pages is just a spreadsheet of guesses.

The output of a proper keyword research project is not a list. It’s a content plan: which pages to build, in what order, targeting which terms, competing against which URLs currently ranking. If your keyword research doesn’t end in a page-by-page plan, you did research, not strategy.

What Are the 8 Steps in a Complete Keyword Research Process?

Every solid keyword research project, whether it’s for a 10-page brochure site or a 500-SKU ecommerce store, follows the same skeleton. The tools and the volume of keywords change; the sequence doesn’t.

1. Define the business goal. Decide what you’re trying to rank for and why — new category, more leads, more transactional traffic — before you open a keyword tool.
2. Build a seed keyword list. Write down 15-30 broad terms your business genuinely competes on, sourced from your own site, sales calls, and product pages.
3. Expand with tools. Run seeds through Google Keyword Planner, Ahrefs or Semrush, autocomplete, and People Also Ask to turn 20 seeds into 500-2,000 candidate keywords.
4. Mine competitors. Pull the keywords 3-5 direct competitors rank for that you don’t, using a gap analysis report.
5. Classify search intent. Tag every keyword informational, commercial, transactional, or navigational — this decides the content format you build, more than the topic itself does.
6. Score difficulty and volume. Check each keyword’s KD score and volume range against your site’s current authority to separate realistic targets from wish-list terms.
7. Cluster and map. Group keywords that share intent into single pages, and assign one primary keyword per URL to avoid two pages fighting each other.
8. Mine Search Console. Check existing rankings for terms sitting on page 2 that need a small push, not a new page.

How Do You Build a Seed Keyword List?

Start with your own site, not a tool. List every product, service, and problem your business solves in plain language — the words a customer would use, not your internal jargon. A 20-minute call with your sales team usually surfaces 10-15 phrases your keyword tool would never guess, because customers describe problems, not solutions.

Add your top 3-5 competitors’ homepage and service-page headlines to the list. Add the categories in your own site navigation. By the end you should have 15-30 seed terms — broad enough to expand, specific enough to be recognizably about your business. We go deeper on sourcing seeds, including a repeatable brainstorming template, in our guide to building a seed keyword list.

How Do You Expand Seed Keywords Into a Full List?

Run each seed through a paid keyword tool (Ahrefs, Semrush, or Ubersuggest) and Google’s own autocomplete and “People Also Ask” boxes. A single seed like “email marketing” typically returns 200-800 related terms once you include modifiers (best, vs, for [industry], near me, how to). Export everything into one sheet before you filter anything — cutting keywords too early is the most common mistake we see in first-time keyword research projects.

Two under-used expansion sources deserve their own workflow rather than a quick mention here: mining Reddit and Quora threads for the exact phrasing customers use when tools haven’t caught up yet, and treating Google’s autocomplete and related-search boxes as a live keyword tool rather than decoration. We cover both in mining Reddit and Quora for keywords tools miss and using Google autocomplete for keyword research.

How Do You Mine Competitor Keywords and Find Content Gaps?

Pick 3-5 competitors who compete for the same customer, not necessarily the biggest names in your industry. Run their domains through a gap analysis tool to find keywords they rank in the top 10 for that you don’t rank for at all. This step alone usually surfaces 30-100 realistic keyword opportunities you would never think to brainstorm, because someone else already validated the demand by ranking for it.

Separate branded competitor searches (people searching a competitor’s name directly) from generic gap keywords — they need different handling and different pages. Our full breakdown of the process, including how to build a repeatable gap report, is in keyword gap analysis, finding competitor keywords, and handling competitor brand keywords.

How Do You Judge Search Intent for Each Keyword?

Search intent is what the searcher wants to happen next, and it decides your content format more than the keyword itself does. Four intent types cover almost every query:

The fastest way to check intent for any keyword: open an incognito tab, search it, and look at what’s actually ranking in the top 5. If four of five results are product pages, Google has already decided the intent is transactional — a blog post won’t outrank them no matter how well it’s written. We break this test down further in search intent analysis, the four search intent types, and telling commercial and informational keywords apart. If you’re specifically hunting for keywords close to a purchase decision, see bottom-of-funnel keyword research.

How Do You Check Keyword Difficulty and Decide What to Prioritize?

Keyword difficulty (KD) scores estimate how hard a keyword is to rank for, usually on a 0-100 scale, based on the backlink profile of pages already ranking. No KD score is a guarantee — it’s a backlink-based proxy, not a ranking prediction — but it’s still the fastest first filter for a list of hundreds of keywords.

Don’t prioritize on difficulty alone — cross it against volume and business relevance. A KD-15 keyword with 20 searches a month and no buying intent is rarely worth a page; a KD-45 keyword that maps directly to your highest-margin service usually is. We lay out a full weighted scoring model, built to replace gut-feel ranking, in how to prioritize a keyword list and go deeper on the score itself in what keyword difficulty scores measure and what they miss.

How Accurate Is Keyword Search Volume Data, Really?

Less accurate than most dashboards make it look. Ahrefs ran a study comparing its own search volume estimates and Google Keyword Planner’s estimates against real Google Search Console impression data for 72,000 keywords. Ahrefs’ numbers landed within 50% of the real GSC figure 60% of the time; Google Keyword Planner’s own numbers did the same only 45% of the time — worse than the tool that pulls straight from Google’s own search data (source: Ahrefs, GSC vs. GKP volume study).

The practical takeaway: treat every volume number as a rough order of magnitude, not a forecast. A keyword showing “1,000-10,000” isn’t 5,500 — it could be 1,200 or 9,000, and both change your content decision. Cross-reference tool estimates against your own Search Console impressions once you have any ranking at all; your own data beats any third-party estimate for your specific site. Full detail on why this happens and how to work around it is in how accurate is keyword search volume, really. Volume also isn’t flat across the year — see our breakdown of seasonal keyword volume swings and don’t write off a keyword just because a tool shows zero, which we cover in zero-volume keywords that still deserve a page.

How Do You Cluster Keywords and Map Them to Pages?

Clustering means grouping keywords that a single page can realistically satisfy, based on shared intent and overlapping ranking pages — not shared words. “Email marketing software” and “best email marketing tools” cluster together because the same comparison page can rank for both; “email marketing software” and “how does email marketing work” don’t, because one wants a product page and the other wants a guide.

Once clusters are set, mapping is the step where each cluster gets assigned to exactly one URL — an existing page you’ll update, or a new page you’ll build. Skip this step and you get keyword cannibalization: two pages on your own site competing against each other in the same search results, which dilutes both instead of ranking either one well. The full clustering method, including a spreadsheet template you can copy, is in keyword clustering and the keyword clustering spreadsheet; the mapping process itself is in keyword mapping.

How Do You Mine Google Search Console for Quick Wins?

Open Search Console, go to Performance, and filter for queries where your average position sits between 8 and 20. These are “striking distance” keywords — you’re already ranking, Google already trusts the page for the term, and a moderate on-page update (better title tag, a missing subheading, 300 more words of real detail) can push you onto page 1 in weeks rather than the months a brand-new page needs.

This is the fastest ROI step in the entire keyword research process, and the one most teams skip because it feels less exciting than building new pages. Our full walkthrough, including the exact filter settings to use, is in striking distance keywords and turning Search Console data into content ideas.

How Long Does a Full Keyword Research Project Take?

Launching keyword research for a brand-new category — no existing rankings, no historical Search Console data, sometimes not even a settled name for the category yet — takes longer and needs a different starting point than refreshing an established site’s list. We wrote a separate process for that exact situation in keyword research for a new category, and a template for turning any of these projects into a deliverable a client or manager will actually read in the keyword research deliverable.

Further Reading: The Full Keyword Research Library

This guide is the hub for our full keyword research cluster. Below is every sub-topic post, grouped by theme, so you can jump straight to the piece that matches where you’re stuck.

Keyword Research Guide (sourcing, tools, and data quality)

• Seed keywords: how to build the list you’ll expand from
• Keyword research with free tools only
• Keyword difficulty scores: what they measure and what they miss
• Mining Reddit and Quora for keywords tools miss
• Question keywords: building a content engine from them
• Zero-volume keywords that still deserve a page
• Google autocomplete and related searches as research tools
• Google Trends for content planning: practical uses
• How accurate is keyword search volume, really
• Turning keyword research into a deliverable people read
• Seasonal keyword volume: reading the swings correctly
• Keyword research for a brand-new category

Search Intent

• Search intent analysis: the incognito-tab method
• The four search intent types, explained with examples
• Commercial vs. informational keywords: telling them apart
• Bottom-of-funnel keywords worth targeting first

Competitor & Gap Analysis

• Keyword gap analysis: the step-by-step process
• How to find the exact keywords a competitor ranks for
• Handling competitor brand keywords the right way

Keyword Mapping & Clustering

• Keyword mapping: assigning one keyword per page
• Keyword clustering: grouping by shared intent, not shared words
• The keyword clustering spreadsheet (with template)
• How to prioritize a keyword list without guessing

Search Console Mining

• Striking distance keywords: the fastest wins in SEO
• Turning Search Console data into content ideas

Vertical Keyword Research

• B2B keyword research: what’s different
• Ecommerce keyword research, product by product
• Local keyword research for city and neighborhood pages
• Branded keyword strategy: protecting your own name

Should You Do Keyword Research Yourself or Hire It Out?

Do it yourself if you have 15-25 hours to spend, access to at least one paid keyword tool, and the discipline to actually finish the mapping step instead of stopping at a raw keyword list. Most founders who try this once realize the research itself isn’t the hard part — turning 800 keywords into a prioritized, page-by-page content calendar is where projects stall.

If you’d rather skip the spreadsheet and get a finished, prioritized keyword map, our one-time keyword research and mapping service starts at ₹500 for single-topic sprints and covers full-site projects too — no custom-quote games, the price is on the page. If you’re not sure whether your current keyword targeting is even the problem, start with the free SEO audit instead.

FAQ

What’s the difference between keyword research and SEO strategy?

Keyword research finds and prioritizes the terms worth targeting. SEO strategy decides what to do about them — which pages to build or update, in what order, with what internal linking and content depth. Keyword research is an input to strategy, not the strategy itself.

How many keywords should one page target?

One primary keyword per page, plus 5-15 closely related secondary terms that share the same search intent. Targeting two unrelated primary keywords on one page usually means neither ranks as well as a focused page would.

Is keyword research still worth doing with AI search growing?

Yes. AI Overviews and chatbots still pull from pages that rank well for the underlying query, so understanding what people search and why remains the starting point. The format of the answer may change; the need to know the exact question doesn’t.

Which free tool gives the most accurate search volume?

None of them are fully accurate — Google Keyword Planner rounds into broad ranges unless you’re running active ad spend, and free third-party tools estimate rather than measure. Google Search Console is the only source showing your actual impressions, but only for terms you already rank for.

How often should you redo keyword research?

Every quarter for an active site — search behavior, competitor content, and your own rankings shift enough in 3 months to change priorities. A full rebuild from scratch is rarely needed more than once a year unless you’re entering a new category.

Written by Palash, founder of PalV’s DM. 5+ years in SEO, 1,000+ articles published.